metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
- name: model_answer
dtype: string
- name: extracted_concepts
sequence: string
- name: reasoning_result
dtype: string
splits:
- name: train
num_bytes: 79091132
num_examples: 867
download_size: 10988380
dataset_size: 79091132
configs:
- config_name: default
data_files:
- split: train
path: data/train-*